SPA-Q: Structure-Preserving Adaptive Post-Training Quantization for Monocular Depth Estimation
cs.CV
Submitted: 2026-11-03
Updated: 2026-11-03
Terminology
Sources
- An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- DINOv2: Learning Robust Visual Features without Supervision
- A White Paper on Neural Network Quantization
- Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation
- FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
- Learned Step Size Quantization
- PACT: Parameterized Clipping Activation for Quantized Neural Networks
- BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstruction
- QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization
- From Big to Small: Multi-Scale Local Planar Guidance for Monocular Depth Estimation
- ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depth
- DopQ-ViT: Towards Distribution-Friendly and Outlier-Aware Post-Training Quantization for Vision Transformers
- DIODE: A Dense Indoor and Outdoor DEpth Dataset
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs